Thermal Insulation with Aluminum Foil RALPH B. MASON, Aluminum Company of America, New Kensington, Pa.
HE principles governing heat flow by conduction, con- vection, and radiation have long been known, but T their application in the production of aluminum foil..
air cell insulation is a relatively recent development. By combining the low thermal conductivity of air with the low emissivity or radiating power of bright aluminum foil in a structure designed to minimize air convection currents, there has been introduced an insulation called Alfol, that l i a ~ met with commercial success (9, If).
Quiet air is one of the hest availilable mediums for prevent- ing heat transfer, and the practical problem resolves itself into a determination of the best structure for confining the air in cells so as to eliminate heat transfer by convection and radia- tion. The most practical structure for any purpose may he a compromise het,ween maximum thermal efficiency and cost.
on aluminum foi la i r eell insulation introduce a variety of experiments1 complications. After a careful survey of avail- able methods, the so-called guarded hot plate and the cold box methods were adapted to the problem a t hand. In the guarded hot plate method, a measured quantity of heat, developed electrically, is forced to flow through the insuls- tion under test. By maintaining a guard ring around the hot plate and at the same temperature, the lateral flow of the measured heat outsidc of the test area is prevented.
The guarded hot plate apparatus was constructed along the lines suggested by Van Dusen (fd) and the National Research Council (7). The faces of the hot and cold plates were made of heavy aluminum sheet. The hot plate, meas- uring 24 X 24 inches (61.0 X 61.0 em.) was so constructed that an inner square on each face, 16 X 16 inches (408 X
The p r e s e n t investigation has determined some of the funda- mental information necessary for the commercial design and utili- zation of this type of insulation.
There are two general nieth- ods of using aluminum foil for insulation. One is to provide a f r a m e w o r k which supports the aluminum foil and forms a series of air cells between the br ight foil s u r f a c e s ; in the other, the foil is first crumpled and then partially s t r e t c h e d so that the resulting wrinkles in the foil separate the sheets when they arc la id a g a i n s t each other and p r o v i d e t h e necessary separation and air cells.
METHong FOR DETERMINING THERMAL CONDUCTIYITY
T h e a c c u r a t e measurement of t h e r m a l conductivity is a difficult determination at hest, and thermal m e a s u r e m e n t s
Thermal insidation with a metal is made possible by laking advantage of the low thermal emissivity of aluminum foil and the low thmnal conductivity of air. The various factors deter- mining the qDiciency of aluminum foil-air cell insulation are analyzed, and the insulation values of a wriety of structwes determined experi- menlally in the guarded hot plate apparatus or the cold bos upparatus. I t is possible with this type of insulation practically to eliminate heal transfer by radiation and convection, and to up- prmch the insulating value of still air. Thp aluminum foil-air cell insulations of the plain air-cell type are found to be better than struc- tures with corrugated separators or crumpled foil. The best results are obtained when the distance between foils is from 0.25 to 0.33 inch (0.64 100.84cm.). The light weight ofaluminum
foil-air cell insulation and its excellent insuh- tion properties make it especially suitable for use in the transportation industry.
40.6 cm.), was isolated from the outer or guard portion by a '/e inch (0.32-cm.) slot. Alternat ing current was used to heat the hot plate, and the energy input WPS determined by means of a Weston wattmeter. The tem- p e r a t u r e s were measured by means of copper-constantan t h e r m o c o u p l e s . A Leeds & Northrup type K potentiometer a n d a sensitive galvanometer were used io conjunction with the tliermocouples.
The test panels were placed in the hot plate apparatus and the wliole assembly well insulated with alumina slag wool and cork board. The conductance was determined at three mean tem- peratures. About 24 hours were usually necessary for the ap- paratus to reach equilibrium for each p o i n t determined. The temperatures were determined by means of the thermocouples and the energy input read from
246 I 1\; D U S T R I A L A N D E N G I N E E R I Iu G C H E JI I S T R Y Vol. 23, No. 3
the wattmeter. All measurements made in the guarded hot plate apparatus and reported here were made upon panels placed in a vertical position in the apparatus.
A I R SPACE WIDTH- INCHES.
CONDUCTAKCE AND WIDTH OF SINGLE ALUMINUM FOIL-AIR CELL INSULA- TIONS AT Two MEAN TEMPERATURES
FIGURE 1. RELATIONSHIP BETWEEN
SPACIKG OF FOIL IN PANELS One of the objects of this investigation was to determine
the optimum spacing for air cells bounded by bright alumi- num foil which would give the maximum heat insulation for a given thickness.
A series of panels was prepared, each approximately 2 inches (5.08 cm.) in thickness, which contained two, three, four, five, six, seven, and eight air cells, respectively, bounded by aluminum foil. The panels were made 24 X 24 inches in area to fit the hot late apparatus, and were similar in construction except for t&e number of sheets of aluminum foil dividing the 2 inches of thickness into air spaces. The sheets of foil were separated and supported by strips of Masonite of the requisite thickness to give the desired number of air cells per 2 inches. The con-
ri . I5
'I ; .14 2 2 .'3 m
L L .I2
II 0 z 0 u .IO
60 80 100 120 I40 IN 180 200 MEAN TEMPT.
FIGURE 2. CONDUCTANCE VALUES OF ALUMI- NUM FOIL-AIR CELL INSULATIONS 2 INCHES
struction of the Masonite separator frames was somewhat com- plicated because of the fact that the guarded hot plate had a measuring area of 16 X 16 inches. The outer shell of each separator frame was made from strips of Masonite 1 inch (2.54 cm.) wide. Crosspieces of the same separator material I/? inch (1.27 cm.) wide were used to form an inner square 16 X 16 inches, which coincides with the hot plate in size and position and is the area of insulation actually under test. This central s uare was divided into four smaller cells, each of which was 7 3 7 4 X 7 3 / 4 inches (19.7 X 19.7 cm.). The purpose of the crosspieces in the center area was to support the aluminum foil and insure good contact with the hot and cold plates in the test apparatus. The aluminum foil was fastened to the separator frames with a rubber resin cement. The Masonite used in the inner square
(16 X 16 inches) or measuring area occupied about 6.15 per cent of that area. Only one height of air cell was measured in these experiments-namely, 7 3 / d inches.
Most measurements of air cell insulation reported in the literature have been made on single air cells. It is much easier to construct panels of this type, but the thermal insulation of narrow air cells is low, and the energy input to the hot plate is small for temperature differences of 10' or 20" F. (5" or 10" C.) between the hot and cold plates. It was also considered best to make the measurements with multiple air cells, since single cells are seldom used.
The first series of measurements with the 2-inch (5.08- cm.) panels but with varying numbers of air cells is shown in Table I. The thermal conductance of each panel was de- termined a t three different mean temperatures. The tem- peratures of the hot and cold plates are reported in the table, as a knowledge of these factors is essential in air cell insula- tion.
In thermal measurements of this type, the common prac- tice is to hold the cold plate constant and increase the tem- perature of the hot plate. It proved to be impractical to hold the cold plate a t the same definite temperature for all the experiments. It was also difficult to construct panels exactly 2 inches in thickness, and because of slight variations
3 Air Spaces
3 4 5 6 7 8 A I R SPACE W I D T H - INCHES
FIGURE 3. RELATIONSHIP BETWEEN CONDUCTANCE OF 2 INCHES OF ALU- MINUM FOIL-AIR CELL INSULATION AND WIDTH OF INDIVIDUAL AIR
in thickness the data are not directly comparable. To make comparisons easy, the data of Table I were first plotted for a fixed cold-plate temperature of 40" F. (4.44' C.). To do this, small corrections were applied to the measured con- ductances in order to adjust the data for a cold-plate tem- perature of 40' F. These corrections were small, being less than the probable error in thermal measurements of this type, except in the case of the insulations containing three air spaces per 2 inches, where there is considerable convection and where differences in temperature play a more important role. From these corrected values the conductances of single air spaces were calculated for mean temperatures of 100' F. (37.8" C.) and 150" F. (65.6" C.) as shown in Figure 1. For example, the average width of each air space in the in- sulation containing six air spaces per 21/8 inches (5.4 cm.) was 0.354 inch (0.899 cm.) (allowance was made for the aluminum foil). The conductance of this insulation a t a mean temperature of 100" F. was 0.100 B. t. u. (0.488 kg.- cal.), and the conductance of a single air cell would be 0.600 B. t. u. (2.929 kg.-cal.) a t the same mean temperature. The relationship between conductance and width of single alumi- num foil-air cell insulations a t mean temperatures of 100' and 150" F. is shown in Figure 1. In drawing smooth curves through the points, a minimum conductance is indicated as
I (2.54) 1 (2.64)
2J/n ( 5 . 3 2 )
I > & / , * (4.82)
. . 85.4 41.2 44.2 03.3 (17.4) 112.3 40.6 71.7 76.5 (24.71 132.3 40.3 92.0 86.3 (30.2)
. . . , , . , . 86.0 130.01
ciccurring fur air spaces betwen 0.6 tmd 0.7 inch (152 and 1.78 cm.) in t,hickncss. Tising approximately the same height air cell, Dickinson and Van Dusen (S) fonnd a mini- mum condiictance at ahout 0.03 inch (1.6 cm.).
Using the curves of Figure 1, a derived series of curves (Figure 2) w ~ s plotted for exactly 2 inches of insulation, and for the condition where the mold face of the insulation is st 40' F. For example, the. conductance of a single air space
inch in width a t 100' F. may be taken from the curve in Figure 1. The conductance of four air spaces (2 inches in total thickness) will be onefourth the conductance of a single air cell. The dotted portion of the curve for the insulation containing three air spaces is uncertain, since the conduo- tance measurements in this region were erratic, as a result, no doubt, of the increased effect of convection. As the num- ber of air spaces in 2 inches of insulation is increased from four to eight, a decrease in conductance is noted. The differ- ence bet&een the conduct- ance of the insulation con- taining seven air spaces and the one containing eight air spaces is small; the one ex- tra foil adds to the insula- tion value by decreasing slightly the small fraction of h e a t t ransmiss ion by radiation.
The optimum spacing for the air cells with a given number of aluminum foils would be about 0.6inch (1.5 em.), as is shown in Figure 1. However, where maxi- mum insulation must be obtained in any givenspace, t h e m i n i m u m conduct- ance is obtained by using more foils with closer spac- ing. Figure3, whichis from the data of Figure 2, shows the relationship between the conductance, at a mean temperature of 60" F , of 2 inches of aluminum foil-air
B . t. zi. (K#.-cd.) 0.142 (0.693) 0.144 (0.703) 0.145 (0.708) 0.302 (1.4741 0.419 (2.046) 0.600 (a.W211) 0.050 (3.173) 0.123 (O.RO0) 0.145 (0.70s) 0.15s (0.771) 0.112 (0.547) 0.117 (0.671) 0.130 (0.035) 0.110 (0.537) 0.1111 (O.SS1) 0.126 (0.615) 0.101 (0.493) 0.10S (0.527) 0.116 (0.566) 0.106 ( 0 . 5 1 7 ) 0.115 (0.561) 0.119 (0.5811 0.103 (0.503) 0.110 (0.537) 0.117 (0.571)
cell insulation and the width oS iridividual air cells. The in- sulation containing eight air cells per 2 inches gave the lowest value in this series OS measurements. If the conductance of 2 inches of insulation containing ten air spaces is calculated from Lhe curves of Figure 1, it will be found to have ap- proxirnatcly t.he same value as the one containing eight. air cells. Ilovever, no measurements were made in this region.
The curve shown in Figure 3 is similar to that obtained by Cregg (4), but the conductance values reported by him Sor aluminum foil-air cell insulation are slightly higher than those given here. However, the material used for the separator Srames in this investigation has a lower conductance than the wooden separator frames used by Gregg, and, if allowance is niade for this difference, the results check in a very satissactory manner. Most of the oonductance values reported by Queer (8) are somewhat hialter than the results
COLD Box APPARATUS
given here. The curves of Figures 2
and 3 show that the in- sulation with the '/Xnch (0.63-em.) spacing gave the lowest conductance of the in su la t ions measured. After allowing for the small amount of heat transferred by radiation, it was found that the ratio of heat trans- ferred to the conductance of still air was practically constant for 2 inches of in- sulation c o n t a i n i n g s ix , seven, and eight air spaces, which mould indicate that the effect of convection was constant and, as will he shown, probably negligible. The experiments of Creregg (4) have shown that the conductances of air cells inch (1.27 em.) in width do not vary, when measured in the vertical position and in a horizontal position. with the
I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y Vol. 25, No. 3
TABLE 11. THERMAL COXDUCTIVITY ASD CONDUCTANCE OF ALUMINUM FOIL AND PAPER INSULATIONS (Plain air-cell type)
TEMPERATURE -. Av. WIDTH TOT.AL THICK-
OF EACH A-ESS OF High Low Dif- PANEL DESCRIPTION OF INSULATION AIR SPACE INSULATION aide side ference Mean CONDUCTIVITY4 CONDUCTANCE
Inch (cm.) Inch (em.) F . O F . O F. O F . (" C.) Center baffled Masonite frames + black 0 . 3 4 (0 .86) 2 .09 (5.31) 105 .1 7 8 . 0 2 7 . 1 9 1 . 6 (33.1)
paper (0.0095-inch or 0.0241-cm.) : 6 1 6 6 . 8 7 8 . 2 8 8 . 6 122 .5 (50.3) air spaces 214 .4 7 8 . 8 135.6 146.6 (63.7)
C. B. Masoniteframes + aluminum- 0 . 3 4 (0 .86) 2 . 0 6 (5 .23) 120 .8 7 2 . 3 4 8 . 5 9 6 . 5 (35 .8) painted paper; 6 air spaces 181 .3 7 3 . 2 108 .1 127.2 (52 .9)
2 5 1 . 4 7 3 . 0 178 .4 162.2 (72 .3) C. B. Masonite frames + 0.0006-inch 0 . 3 4 (0 .86) 2 . 0 6 (5 .23) 133.2 7 6 . 7 5 6 . 5 104.9 (40.5)
(0.0015-c,m.) lacquered foil (thin coat- 198 .3 7 6 . 2 122.1 137 .3 (58.5) ing) : 6 air spaces 276 .5 7 6 . 1 2 0 0 . 4 176 .3 (80 .2)
Masoniteframesfromcoldbox + 0.00035- 0 . 4 4 ( 1 . 1 2 ) 0 . 8 8 (2 .24) 109.2 4 0 . 2 6 9 . 0 7 4 . 7 (23.7) inch (0.00089-cm.) plain foil: 2 air 176.4 4 1 . 3 135 .1 108.9 (42 .7) spaces 234.6 4 1 . 8 192.8 138.2 (59 .0)
Masonite frames from cold box + 0.00035- 0 . 4 4 (1 .12) 0 . 8 8 (2 .24) 9 7 . 2 4 2 . 0 5 5 . 2 6 9 . 6 (20 .9) inch lacquered foil: 2 air spaces 155 .8 4 2 . 5 113 .3 9 9 . 2 (37 .3)
204 .0 4 2 . 0 162 .0 123 .0 (50.6)
B. t. U. (Ku.-cu~.) B. t. U. ( K g . - ~ d . ) 0.470 (0.0583) . . . 0.510 (0.0632) . . . ... 0.552 (0.0684) , . , 0 . 2 7 0 (0.0335) . . . . . . 0 . 3 1 1 0.0386) . . . ...
0 . 3 3 4 10.0414) ... ... . . . . . . ...
0.217 (0.0269) 0 .247 (1.206) 0 .246 (0.0305) 0 . 2 7 9 (1.362) 0 . 2 7 5 (0.0341) 0 .312 (1.523). 0 . 2 5 6 (0.0317) 0 .291 (1.421) 0 .295 (0.0366) 0 .335 (1.635) 0 .328 (0.0407) 0 .373 (1.821)
a All values for thermal conductivity are in B. t . u./hour/sq. ft /" F./inch of thickness, and kg.-cal./hour/sq. meter/O C./meter of thicknesa.
TABLE 111. DESCRIPTION OF IXSULATIONS WITH CORRUQATED ASBESTOS SEPARATORS >fAX. TOTAL THICKNESS THICKNESR
PANEL COABTRUCTION OF INSCLATION~
WIDTH THICKNESS OF OF PITCH OF OF EACH O F CORRUQATED RADIATION CORRU-
BIB SPACE INKLATION SEPARATORS SEIELD QATIONB WEIQET Lb./
Inch (cm.) Inch (cni ) Inch (cm.) Inch (em.) Inch (cm.) cu. ft. (Q./cc.)
I/ls-inch (0.48-cm.) corrugated asbestos separators + l/r-inch (0.64-om.) corrugated asbestos separators + '/a-inch (0.95-cm.) corrugated asbestos separators + '/pin& corrugated asbestos separators + aluminum l/pinch (1.27-cm.) corrugated asbestos separators + l/pinch corrugated asbestos separators + aluminum
asbestos sheet; 5 air spaces 0 . 1 6 (0 .41) 1 . 0 0 (2 .54) 0 . 0 1 8 (0.046) 0 .022 (0.056) 0 . 4 8 (1 .22) 12 .3 (0.197)
aluminum foil: 5 air spaces 0 . 2 1 (0 .53) 1 . 1 3 (2 .87) 0 . 0 1 8 (0.046) 0.0006 (0.0015) 0 . 5 0 (1.27) 7 . 3 (0.117)
asbestos sheet; 3 air spaces 0 . 3 5 (0 .89) 1 . 1 9 (3 .02) 0 .018 (0.046) 0 .018 (0.046) 1 . 2 5 (3.18) 6 . 7 (0.107)
foil; 3 air spaces 0 . 3 3 (0 .84) 1 .06 (2 .69) 0 . 0 1 8 (0.046) 0.0006 (0.0015) 1 . 2 5 (3 .18) 4 . 5 (0.072)
asbestos sheet; 2 air spaces 0 . 4 6 (1.17) 1 . 0 0 (2.54) 0 . 0 1 8 (0.046) 0 . 0 1 8 (0.046) 1 .06 (2.69) 6 . 4 (0.102)
foil; 2 air spaces 0 . 4 8 (1.22) 1.00 (2.54) 0 . 0 1 8 (0.046) 0.0006 (0.0015) 1 .06 (2.69) 4 . 4 (0.070) 4 Sodium silicate was used as adhesive: insulations were baked before measuring thermal conductivity.
source of heat either above or below; this would indicate that there was no convection present. The value given in the International Critical Tables for the conductance of air is a mean value. When the somewhat higher values for the conductance of air, as given in the literature, are used, and when corrections made for the conductance of heat by the Masonite strips included in the measuring area are con- sidered, it is found that the conductivity of still air has been very closely approximated in this form of insulation. Con- sidering the various factors involved, it appears that very little could be gained by using closer spacings than inch in the aluminum foil-air cell form of insulation.
When the width of the air space exceeds ' / z inch, convection starts to play an important role, and the greater the thermal head (temperature difference) between the foils, the greater will be the amount of heat transferred at a given mean tem- perature.
EFFECT OF EMISSIVITY ON CONDUCTIVITY
To demonstrate the part played by radiation in heat transfer, various mate- rials such as black paper, aluminum- painted paper, and lacquered foil were substituted for the aluminum foil in the Masonite frame panels containing six air cells per 2 inches. The results are shown in Table I1 and Figure 4. For purposes of comparison, the results obtained with the bright foil (D) are also included. The panels containing foil covered with a very thin coat of lacquer tC) were only slightly inferior to the bright foil, showing that the emissivity of the foil
had been only slightly increased by lacquering. The conductance of the panels constructed with aluminum- painted paper ( B ) was approximately what one would expect with a material having an emissivity of about 30 to 40 per cent. The substitution of black paper ( A ) for the foil decreased very markedly the heat-insulating properties of the air cells. White paper gave results which were prac- tically the same as those obtained with the black paper, show- ing that the long heat rays are absorbed and emitted as readily by the white paper as by the black.
Another comparison between bright foil and lacquered foil is shown in Figure 5. In this case a heavy coat of protective lacquer was applied to the foil, and the conductance has been increased more markedly (about 25 per cent) than in the case where there was only a thin coat of lacquer. These curves were not included in Fievre 4. since the seuarator
ALUMINUM FOIL-AIR CELL PANEL Type measured in hot plate apparatus;
outer foil removed to show inner construc- tion of panel.
frames were of different consiruction and were only 0.88 inch (2.24 cm.) in thickness. The separator frames were made from Masonite strips, but instead of isolating an inner square 16 X 16 inches (40.6 X 40.6 cm.) from the 24 X 24 inch (61.0 X 61.0 cm.) frame, cross- pieces inch (1.27 cm.) wide were used to form nine air cells approximately 7 x 7 inch (17.8 X 17.8 cm.). When these frames were placed in the hot plate apparatus, the air cells overlapped the guard and test portions of the hot plate. The results plotted in the usual way indicate that steeper curve8 Oligher temperature coefficient) are obtained where this condition exists.
Figure 4 shows that a very thin coat of lacquer on the aluminum foil'
March. 1933 I N D U S T R I A I. A N 1) E N G I N E E R I N G c 11 E M 1 s'r H Y 249
GUARDED HOT PLATE. APPAEUTW
did not appreciably increase tile coiiductivity of the ilk- sulation in which the foil was incorporated. A very tliick lacquer coating, however, resulted in a larger in- crease in conductance (Figure 5). These facts are of im- portance where the insulation may be used under corrosive conditions and where a thin coating of lacquer wilt maintain the brightness and low emissivity of the foil and yet give satisfactory protection against corrosion.
PANELS WITH CORRUGATED SEPARATolLs
For many purposes there are advantages in the use of thin corrugated asbestos or paper structures separating the aluminum foils. A series of panels was constructed in which the foil layers were separated by corrugated asbestos, and,
p 46 i 56 t 5 2
z :; 32
E .28 2 .21( 0 z 0 sa iw i zo 140 so 180 0
M L A N T I M P ' f .
FIGURE 4. EFFECT OF E M ~ I V I T Y 01' AIR-CELL PhanTioNs ON C o ~ ~ u c ~ t v r n (SIX Am SPACES PER 2 INCHF-R AWIIOXI-
for comparison, panels were iiicludcd in wlrich asbestos sheel WES substituted for the aluminum foil. A brief description of a e h set of panels is given in Table 111. The depths of corrugation in the original separators are given in Tablo 111, but, during assembly and clamping into the hot plate, the separator is compressed somevhat and the width of the air cell is changed. The table contains the total thickness of
tlie ~iaiiei, the iiurnber of Njr cells, and the thickness of the rndiatiori shields, and from these data. the maxkium width of each air cell was computed. The results of tests on these panels are shown in Table IV and Figure 6.
The substitution of aluminum foil for the asbestos sheet used between the corrugated separators causes a marked de- crease in the thernial conductivity. The panels with the '/&ch (0.63-cn1.) corrugated separators were not as ef- ficient insulators as the panels with t.he 3/&ch (0.95-em.) corrugated separators. Asbestos has a higher thermal con-
ductaiicc tlran air, and considerable heat passes through the separaton. The presence of more asbestos in the '/&nch corrugated structure than in the 3/s-inch structure probably explains the poorer insulating value. The panels containing the '/r-inch corrugated separators were measured with the corrugations in a vertical position and also with the corn- Rations in a horizontal position. S o difference was noted in tlie conductivity values for the panels as measured in the difiercnt positions. Since asbestos is E fairly poor beat in- sulator, more heat than usual passed through these panels. On account of this, no conclusions will be drawn from these measurements as to the relative merits of vertical and hori- zontal corrugat,ions.
Panels with various types of corrugated papor separators were constructed and are described in Table V. The thermal conductivities o i these insulations are shown in Table VI and Figure 7. In general, the panels constructed with GOITU- gated paper separators were somewhat better thermal imu- lators than those containing corrugated asbestos separators. The paper had a higher thermal resistance and was con- siderably thinner and lightor in weight. 111 addition, the
250 Vol. 25, No. 3 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
TABLE IV. THERMAL CONDUCTIVITY OF AIR-CELL INSLJLATIONS WITH CORRUGATED ASBESTOS SEPARATORS POSITION OF CORRUGATIONS MAX. WIDTH TOTAL THICKNESS I TEMPERATURE
IN OP EACH OF Hi h Low PANELS HOT PLATE APPARATUS AIR SPACE INSULATION si8e side Difference Mean CONDUCTIVITY
Inch (ern.) Inch (em.) O F . O F. O F. F. (" C.) B. L. w. (Kg.-caZ.) 0.459 (0,0569) A Horizontal 0.16 (0.41) 1.00 (2.54) 7 0 . 5 40 .9 29.6 55.7 (13.2)
148.4 40.7 107.7 94.6 (34.8) 0.495 (0.0614) 222.4 41.8 181.1 131.9 (55.5) 0.528 F .0655)
B Horizontal 0 .21 (0.53) 1 .13 (2.87) 9 9 . 4 48 .6 5 0 . 8 7 4 . 0 (23.3) 0.365 0.0453) 198.5 60.0 148.5 124.3 (51.3) 0.384 0.0476)
0.400 0,0496) 276.1 50.4 225.7 163.3 (72.9) B Vertical 0 .21 (0.53)
C Horizontal 0 .35 (0.89)
D Horizontal 0 . 3 3 (0.84)
E Horizontal 0.46 (1.17)
F Horizontal 0.48 (1.22)
.13 (2.87) 105 .8 5 6 . 5 49 .3 81.2 (27.3) 0 .373 (0.0463) 198.8 5 6 . 8 142.0 127.8 (53.2) 0.394 (0.0489) 262.2 58 .9 203.3 160.5 (71.4) 0 .405 (0.0502) . .
.19 (3.02) 84 .6 40 .9 43.7 62 .7 (17.1) 0.443 (0.0549) 200.5 40 .7 159.8 120.6 (49.2) 0.534 (0.0662) 255.7 41 .4 214.3 148.5 (64.7) 0 .580 (0.0719)
0.298 (0.0370) 222.3 42.9 179.4 132.6 (55.9) 0.325 (0.0403) 311.8 44.4 267.4 178.1 (81.2) 0.348 (0.0432)
.06 (2.69) 104.0 42 .5 6 1 . 5 7 3 . 3 (22.9)
7 5 . 1 164.2 251.0 83 .6
4 1 . 1 41.4 41.3 41 .3 41.9 41.6
34.0 122.8 209.7 42.3
58.1 (14.5) 102.8 (39.3) 146.2 (63.4) 62.5 (16.9)
121.4 (49.7) 149.4 (65.2)
0.482 (0.0598) 0.581 (0.0720) 0.688 (0.0853) 0.386 (0.0479) 0.450 (0.0558) 0.476 (0.0590)
Described in Table 111.
TABLE V. DESCRIPTION OF INSULATIONS WITH CORRUGATED PAPER SEPARATORS Max. TOTAL THICKNESS THICKNESS PITCH
WIDTH THICKNESS OF OF OF OF EACH O F CORRUGATED ALUMINUM CORRU-
PANEL CONSTRUCTION OF INSULATION" AIR SPACE INSULATION SEPARATORS FOIL GATIONS WEIGHT
E F G
'/,-inch (0.64-om.) corrugated kraft paper separa- tors + aluminum foil; 9 air spaces
6/1o-inch (0.79-cm.) corrugated paper separators + aluminum foil. 7 air spaces
*/a-inch (0.95-cd.) corrugated kraft paper separa- tors + aluminum foil: 6 air 8 aces
:/a-inch corrugated chestnut Zber separators + aluminum foil. corrugations crossed and sprayed with alurninuA paint. 6 air 8 aces
s/s-inoh corrugated chkstnut h e r separators + aluminum foil. 6 air spaces
*/a-inch corruga'ted chestnut fiber separators + aluminum foil. crossed corrugations: 6 air spaces
'/a-inch corruiated strawboard separators + aluminum foil: 6 air spaces
'/%-inch (1.27-cm.) corrugated kraft paper separa- tors + aluminum foil: 4 air spaces
idium silicate was used as adhesive except for panel C
0 .27 (0.69)
0 .28 (0.71)
0 .32 (0.81)
0 . 3 1 (0.79)
0 . 3 0 (0.76)
0 . 4 3 (1.09) p , where rubt
1 . 7 5 (4.45)
1 .72 (4.37)
2 .00 (5.08)
1 . 9 1 (4.85)
1 .97 (5.00)
1.72 (4.37) ,er cement was
panels with the 3/Anch corrugated chestnut fiber separators were tested not only with the corrugations running in a horizontal, but also in a vertical, direction. Table VI gives also the results of these measurements, which indicate that the higher thermal resistance is obtained with the corruga-
s 72 \ 66 2 64
= 5 :; ? 48 544 5 40 2 36
V g 32
M E A N TEMP 'F FIGURE 6. CONDUCTIVITY OF ALU- MINUM FOIL INSULATIONS WITH COR-
RUGATED ASBESTOS SEPARATORS
F. '/t-inch separators + aluminum foil tions running in a vertical direction. However. it was noted in these measurements that the difference between the tem- peratures a t the top and bottom of the insulation was several degrees when the corrugations ran in a vertical direction,
0.0087 (0.022 1)
0.0029 (0,0074) used.
1 . 2 5 (3.18)
1 . 2 5 (3.18)
1 . 2 5 (3.18)
Lb./ cu. ft.
1 . 9
1 . 9
1 . 3
2 . 8
2 . 5
2 . 2
1 . 0
whereas it was only a fractional part of a degree when the corrugations ran in a horizontal direction. This greater difference in temperatures was also noted when panels with crowed corrugations were tested.
Spraying the corrugated separators with aluminum paint improved the thermal resistance of the insulation panels. As would be expected, the corrugated separators with the highest thermal resistance made the best insulation panels. The insulations with the 3/8-in~h corrugated paper separators are better than those with the '/(-inch corrugated paper separators. This was also found to be true with the corru- gated asbestos separators. The results obtained with the corrugated paper separators showed that they were not as good thermal insulators as the straight air-cell type. This difference will be discussed later.
CRUMPLED FOIL INSULATION For some purposes the crumpled foil type of insulation
has advantages, notably where it must be installed over and around curved and irregularly shaped surfaces. In measur- ing the insulation value of crumpled foil, special frames were constructed from Masonite strips in order t o hold the crumpled foil in place between the hot and cold plates of the test apparatus. The standard procedure was to reduce the foil 10 per cent in length by crumpling with the hands; in one test, however, the foil was reduced 23 per cent in length to show the effect of this variation. The results of tests on a variety of panels are shown in Table VI1 and Figures 8 and 9. The thickness of the aluminum foil employed did not appear to influence the thermal conductivity appreciably,
March, 1933 I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
TABLE VI. THERMAL CONDUCTIVITY OF ALTJMINUM FOIL INSULATIONS WITH CORRUQATED PAPER SEPARATORS
POSITION OF CORRUGATIOX~ I N HOT P L h T E APPARATUS
Vertical; baffled a t 16 inches (40.6 om.)
a Described in Table V.
L f A X . WIDTH O F EACH
AIR SPACE Inch (cm.)
0 .27 (0.69)
0 .28 (0.71)
0 .31 (0.79)
although the thinnest foil gave slightly the best results. There is also no very marked difference in the conductance values obtained with similar panels fabricated with either plain or embossed foil. In comparing the effect of the de- gree of crumpling, a higher thermal conductance was ob- served in the case of the panel made with foil reduced 23 per cent in length, as compared with other panels in which the foil was reduced 10 per cent in length by crumpling.
M E A N TEMP'F. FIGURE 7. CONDUCTIVITY OF ALU- MINUM FOIL INSULATIONS WITH COR-
RUGATED PAPER SEPARATORS A. I/r-inch corrugated kraft paper separators
aluminum foil t i .+s/~einch corrugated paper separators +
aluminum foil C. a/s-inch corrugated kraft paper separa-
tors + aluminum foil D. a/s-inoh corrugated chestnut fiber sepa-
rators (coated with aluminum paint) + aluminum foil
E. a/s-inoh Corrugated chestnut fiber separa- tors + aluminum foil: horizontal corruga- tions
F . a/s-inch corrugated chestnut fiber separa- tors + aluminum foil; crossed corrugations
G. 8) s-inch corrugated strawboard separators + aluminum foil H . '/*-inch corrugated kraft paper separa-
tors + aluminum foil
I n Figure 9 are shown the results obtained with a 2-inch crumpled foil insulation with six and seven layers of foil (curves A and B, respectively). From the slope of the
TEMPERATURE High side O F.
111.2 163.6 229.1 111.1 167.8 224.1 114.0 171.8 231.3 114.6 182.0 228.5 118.6 173.5 260.3 123.0 176 .8 239.0 123.4 178.9 234.0 119.3 167.2 221.8 121.0 183.3 243.6 110.8 157.6 218.9
O F. 7 0 . 1 7 0 . 4 70 .7 67.8 67 .2 66.6 7 2 . 1 71.6 70.9 64 .0 64 .6 65.0 74.9 75 .4 76 .4 77.2 76.9 76.2 76 .3 7 7 . 1 78 .0 7 4 . 3 74 .8 74 .9 7 4 . 5 7 5 . 8 76 .6 75 .2 74.7 74.6
4 1 . 1 93.2
100.6 157 .5 41.9
100.2 160.4 50.6
117.4 163.5 43.7 9 8 . 1
183.9 45 .8 99 .9
162.8 4 7 . 1
101.8 166.0 45 .0 92 .4
146 ,9 46 .5
107.5 167.0 35.6 8 2 . 9
a F. 90.7
117.0 149.9 8 9 . 5
121.t 151.1 8 9 . 3
123.2 146.7 96.7
124.4 168.3 100.1 126 9 157.6 99.9
128.0 166.0 96 .8
121.0 148.4 9 7 . 8
129.5 160.1 93 .0
Mean (" C.) (32,6) (47.2) (65.5) (32.0) (47.5) (63.0) (33.9) (49.8) (66.2) (31.8) (50.7) (63.7)
$2. ;; (75.7) (37.8) (52.7) (69.8) (37.7) (53.3) (68.9) (36.0) (49.4) (64.7) (36.6) (54.2) (71.2) (33.9) (46.8) (63.7)
CONDECTIVITY E . t . u . 0.268 0 ,287 0.310 0.275 0.303 0.324 0,255 0.277 0.286 0.254 0.280 0.300 0.275 0.288 0.308 0.271 0.279 0.290 0.264 0.277 0 ,296 0 ,259 0.294 0 ,333 0.262 0.268 0.290 0.313 0.335 0.361
(Kg.-cal.) (0.0332) (0.0356) (0.0384) (0.0341) (0.0376) (0.0402) (0.0316) (0.0343) (0.0355) (0.0315) (0.0347) (0.0372) (0.0341) (0.0357) (0,0382) (0.0336) (0.0346) (0.0360) (0.0327) (0.0343) (0.0367) (0,0321) (0.0365) (0.0413) (0.0325) (0.0332) (0.0360) (0.0388) (0.0415) (0.0448)
curves, one would judge that the addition of the extra layer of foil had decreased the convection, as well as improved the insulation values a t the higher temperature. Curve C shows the conductance of 2 inches of crumpled foil insulation calculated from the results obtained for one inch of insulation containing three layers of crumpled foil. The number of aluminum sur- faces exposed to air would be the same in curves B and C.
In making the measure- ments with crumpled foil, no attempt was made to isolate an inner square 16 X 16 inches (40.6 X 40.6 cm.), as was done for the plain air-cell type of in- sulation. The convection currents were free to flow past the guard and meas- uring areas. KO great ac- curacy is claimed for these experiments, and it is to be understood that the hot and cold p l a t e s were made of bright aluminum. It is felt that variations between the d i f f e ren t curves for crumoled foil are
80 100 120 140 160 M E A N TEMP.'F,
FIGURE 8. CONDUCTANCE OF ONE-IR'CH CRUMPLED FOIL IN- SULATION (THREE LAYERS PER
INCH) A. 0.00035-inch laoquered Soil (thin
B. 0.00063-inch embossed foil: 10%R C . 0,00035-inch embossed foil; 10%R D. 0.00035-inch lacquered foil (thin
E. 0.0003'5-inch0 plain foil: 10% R F. 0.00028-inch plain foil; 10% R
coating)' 23% R1
coating)' 10Y R
1 R = reduction in length by crum- pling more apt to beA caused by
the unavoidable difference in crumpling than by the thickness of the foil, the embossing, and the emissivity of the surface.
DISCUSSION OF CONVECTION CURRENTS
The data show that the plain air-cell type of insulation bounded by aluminum foil can be constructed to have a ther- mal resistance approaching that of still air. The effect of the aluminum with its very low emissivity value is well under- stood, and it is readily shown that a relatively small number of radiation shields will decrease the heat transferred by radia-
252 1 R D U S T 1% I A 1, A N 13 E N G I R li Is K I N G C 3 E M 1 S T H Y Vol 25, No. 3
i)s*cil,l.r*"N or INecL,,Tros
0.00035-iech (0.00089-cm.) embossed crumpled f u i l ; 3 h~sers; 10% I
The thermal conductance of the baffle plate must a h he considered.
In the case of a corrugated separator between a hot and a cold surface, the heights of the air cells will be small and the velocity of the convection currents per unit area will be greater. There is a large number of paths of flow directly from the cold surface to the hot surface, and also the thermal conductance of the separator will add materially to the heat, transferred. In an air cell of this type which is bounded by aluminum foil, there is stiU another factor to be considered. The corrugations will cover up a certain u e a of the highly heat-reffecting surface of the aluminum, and any adhesive smeared on the surface will tend to decrease its reflectivity. A consideration of all of these factors explains the lower heat resistance of tho corrugated structure, m d the fact that the '/-inch (0.95-cm.) corrugated separators were better than the L/,-inch (0.63-em.) corrugated separators.
One experiment with a corrugated structure indicated that the insulation hsd a higher resist.ance when the corrugations ran in a vertical direction than when they were horizontal. It will be necessary to carry out several more experiments before this can be definitely proved.
What has been said in regard to the corrugated structures applies also to the crumpled foil insulation. Here there are large irregular air cells which act 89 baffies to the convection currents and offer many paths of Row: Also, some heat is conducted directly through the insulation by the metal foil.
EXPERIMENTS WITH GOLO Box 't'he cold box was desiuned to tnke tive ~anela. 24 X 24 inches
. .... ~.~ ~ ~ . ~ ~ ~ y ice and solid c&bon dioxide were used rn re&iger&ts. The heat input was determined by weighing the water which drained from the inner can when ice was used as the refri ermt and by measuring, b means of a precision gas meter, &e volume of carbon dioxidre gas evolved when solid carbon dioxide was used aa the refrigerant. The teat panels were sealed into the open- ings of the cold box apptlrstus and thermocouples fastened to each face of the inadation. When equilibrium waa reached, the tem- peratures of the inner and outer faces we% determined, as well &s the heat input. The temperature of the room in which the measurements were made was kept fairly Constant by B thermo- stat on the steam radiator. The cold box NI~S first standardized with cork board of known thermal conductivity.
The difficulties encountered in making measurements with the cold box type of apparatus have been described by others, and with them the present author agrees.
The results obtained with ice as the refrigerant are given in Table VIII. The separator frames for the plain air-cell panels were made from Masonite strips. Each frame was 24 X 24 inches, and was divided into nine air cells, ap- proximately 7 X 7 inches (17.8 X 17.8 em.), by cross pieoes
PAPER BACKINO MAY BE APPLIED BY RUNNINO THE ForL ABD THE P A P E ~ Tnnoua TEE ROLL
AT TWE SAME TIME
of '/2-incli (1.27-em.) MasoNte. The corrugated structures were similar to those used in the hot plate apparatus. The results obtained a t lower temperatures vhen Dry-Ice W&L. used as the refrigerant are given in Table IX.
The results obtained on the cold box apparatus appeared better than might have been expected. Pancl2 (Table VIII) WBS tested in the hot plate apparatus, and the results checked in a very satisfactory manner with the results obtained on
Ilurch, 1YJ3 I N L, 11 S 7'1% 1 A 1, A N D I" & G I N E b: 1I I N G C 1 E &I I S T 1% Y 255
easy t o a p p l y a n d can he used to insu la te i r regular - slisped objects for which it is i m p r a c t i c a l t o c o n s t r u c t frrmes for the uniform spacing of the foil. Tho crumpled foil iirsulation weighs very little as compared xi th the other com- mon i n s u l a t i n g m a t e r i ais . The liglrtncss a n d ease of aiq)lication of c rumpled foil are commercial factors which fayor its use i n s t e a d of the plnin air-cel l or corrugated s e p a r a t o r type of aluminum foil i n su la to r s which are rnmewhat more efficient.
tion of the air-cell type. The results obtained show that by careful d e s i p it is possible to construct an insulation from foil in which there is little or no convection. By using the minimum amount of separator material of high thermal resist- ance, the foil insulation can be made t.o approximate the thermal resistance of quiet air. The sec re t of this outstand- ing performance of aluminum foil insulation lies in the low t h e r m a l e m i s s i v i t y of the bright aluminum surface.
The conductivities and ~nuntpi.En ~ " 1 1 . t'aovir,Es .m ~ w m m r ~ m u u n m ACKNOWLEDGMENT xities of various in su la t ing MATE~IAI. FOR T m i w r E m o K oF nIls MEAT ~ n o c x , m a t e r i a l s and the different OPERATBI) BY THE W ~ L L ~ M ZOLLBR COXPANY OF In carrying out this investi- types of a l u m i n u m foil-air PITT~BOBOEI gation the author had the ad- ce l l i n su la t ions a r e snm- vice and assistance of many of marized in Table XI. his associates iii the Aluminum R.esearch Laboratories.
Spocial acknowledgment is extended to Ralph B. Derr, who planned and supervised the construction of the panel insula- tions, and to Cyril S. Taylor, who collabora.ted on the therm1 measurements.
SUMMABY ( I ) Aluminum foil insulations of the plain air-cell type and
approximately 2 inches (5.08 em.) thick have been tested in a guarded hot plate apparatus; the most efficient structures were found to approach the conductanm of still air. Thc
inch (1.69 to 0.63 em.). The insulation which gave the highest resistance to the flow of heat was the one which had
distances between foils in these tests varied from to 1/( LIlEIIATlJliE CITED ( I ) ~ m s r i a m i Soo. Ileatinp and Ventilating Engrs.' Guide, 1932. (a) Am. Soo. Refrigerat,in6 Engrs., Circ. 1, 3rd ed. (1930).
the '/&-inch spacing, The insulation with a distance of 1/8 (3) DiCkinsOn. H. C., and Van Uuuen, M. S.. Am. Soc. Re/r*e~ating Envs. J. , 3, 5--25 (ISIF). inch between foils was the poorest insulation, and, judging
transmitted by convection. For spacings less than inch London, Special Repl. 9 (Oct., 1'322).
rected for radiation) and the conductance of still air was prac-
(4) erCgp, J. I.., ~ ~ , + ~ ~ ~ i ~ ~ E ~ ~ ~ . , 23, 279-83 (1932). from the temperature coefficient, considerable heat was
tically constant. Aluminum foil insulations with corrugated separators
between the sheets of aluminum foil were slightly inferior to the plain air-cell type. The baffle effect of the corrugated separator upon the convection streams, causing cold air to be conducted by many paths directly from the cold to the hot surfaces, the conductance of the separat.or itself, and the covering Of a portion Of the surface, have been suggested as factors which cause a poorer insulation against t.he flow of heat.
were somewhat inferior to the corrugated structures. The amount of heat transmitted by convection appeared to be greater. Here again, there are a large number of baffles UIRLIOGKAPHY to deflect the stroarns Of air, and eonvoction
(1) Coblentr, W.. and Hughes, C. W., Bur. Standards. Tech. Paper streams apparently flow dircctly from t,hc hottest foil to tho coldest foil around the edges, unless special precautions are ( 2 ) Dny, V. S., Univ. Ill. Expt. Sta.. Bull. 117 (1020). taken to prevent this. Also heat may be conducted from one (3) NiohoUa, P., J . Am. SOC. Ileatinu Voritilvtinu Bnurs.. 27. 783
(4) PQclet, "Trail6 do la chdoiir." Vol. I, p. 519 (1878). foil to the next directly through the points of contact. (4) A typical cold box apparat11s was used as a check upon
the results obtained by the guarded hot plate apparatus. The results which were obtained showed good agreement (6) Soh*& L. W., Heatinu, Piping, Air Coditioninu, 2. 957-62
(7) Sohmidt, 13.. "WAmestrablune technisoher ObcrKiichen bei between the two methods of testing. It was also possible to obtain a low mean temperature in the cold box apparatus eew5bnlioher Tomperatur," Reihofte sum pesundheits. w-ith solid carbon dioxide as the refrigerant. The determina- Ingenieur. Seriea 1. No. 20 (1927). tion of the rate of evolution of carbon dioxide gas by means of (8) SInoluchowskr, Maryan von. Proc. bnd RdriuoTatinu Conur.,
1910, 187-93. a gas meter is a novel method for the measurement Of '*le (9) Taylor. T. S., Am. Soo. Refrigerating Engrs., Circ. 1, 3rd ed. rate of heat input to the cold box. (1930); J . Am. Soc. Hedinu Venlilntinu Bnurs., 26. 445 Aluminum foil makes an excellent radiation shield, (1920); Am. SOC. Relriueraliny ETLPS. J . , 10. 179-93 (1923).
and with aluminum foil it is easy to eliminate practically aU of the heat ordinarily transmitted by radiation in an insula-
( 5 ) Gristha, Euor, and Davis. A. H., Food Inwastipation Boaid,
(6) Mull, W., and Reiher, E., "Der W:irnieschutz von Luftliohioh- tcn." Beilefte nuin geaundheits-Ingenieur, Seriw 1, No. 28 (1'330).
(7) Natl. Research Council, Rept. of Comm. on Standard Test Code for Determining Thermal Conduotivity of Boilding Ma- teriiils and Wall Construction. New York. 1929.
(8) Queer, E. R.. Ifealinu, Pifling, Air Cmditiuninu, 3, 960-5 (1931).
(9) Schmidt. E., Z. Ver. dcvi. Iny. , 71, 1395-1400 (1927); u. s. Patent 1 ,8 '~o .m (DSC. e, ieaa).
(10) Sohmidt. E., and Beckmann. W., Z, Ver. dent. I W , 75, 227 (Isal).
(11) Schmidt, E.. and Dyckerhoff, R . , British Patents 266,177 (Feb. 24. !927). 310.572 (MW 2, 10291, and 317,678 (nug. 22, 1929); u. s. Patent 1,757,479 ( ~ a y e, 1030).
the ratio between the measured conductance
(3) The insulation containing foil (12) V*n DWeiL, M. s., J . Am. 8 O C . f i e d < W VWLtihtiW Enym.. 26, 625-56 (1920).
254, 171-87 (1924).
BWT~., 35, 17-26 (1929). (5) R ~ ~ ~ I ~ ~ , F. B., and ugren, A. u.. J . nm. see. ti^^
R ~ ~ ~ , ~ ~ ~ J~~~~~ 12, iw.